FixField Advancing Weed Detection System using Convolutional Neural Network Integrated with IoT Device

In this work, we propose an advanced weed detection system by hybridization of IoT and ML methods. Unlike the formerly used techniques, which almost purely depend on edge devices, we have introduced an advanced object detection framework called YOLOv10, into the proposed system, which sends video analysis back to the cloud while increasing precision and scalability. We utilize an ESP32 camera mounted on agricultural equipment such as tractors for real-time field observation through live video streams. Once the scanning process is complete, this video is sent to the cloud. YOLOv10 then picks and pinpoints weeds in that area. This weeds detection system alerts agricultural practitioners through mobile applications or SMS to intervene immediately. It is therefore cost-effective and very applicable in cases of both large-scale and small-scale farming because instant identification of weeds with minimal human effort is what it offers. Our system truly outperforms what is applied today and has practical and real-world applications toward the achievement of agricultural efficiency together with sustainability in farming practices.